233

Abstract

Listen

Translate article

Introduction: No consensus exists regarding the definition of ventilator-associated pneumonia (VAP). Even within a single institution, variable diagnostic criteria result in conflicting rates of VAP. As a level 1 Trauma center participating in both the Trauma Quality Improvement Project (TQIP) and the National Healthcare Safety Network (NHSN), our institution showed inconsistencies in the rates of culture-based definition compared to TQIP/NTDB. The purpose of this study was to compare rates of VAP determined by our culture-based (Cx)definition to rates determined by the criteria of National Trauma Data Bank (NTDB) and NHSN. Methods: Consecutive trauma patients admitted from January 2008 through December 2011 were included. The National Trauma Registry of American College of Surgeons (NTRACS) and two institutional databases were queried for the presence of VAP in this population, and patients were classified based on which set of criteria they met. Our institution’s criteria to diagnose VAP begin with clinical suspicion (one or more of the following: fever, leukocytosis, purulent sputum, CXR findings, and/or deteriorating pulmonary function). Clinical suspicion triggers deep respiratory culture and the presence of a positive culture (>10k cfu speciated bacteria) defines VAP. Institutional databases identified those meeting NHSN & Cx criteria. Individual chart review was performed to identify those patients meeting NTDB criteria for pneumonia after ≥ 2 days of intubation. Results: In total, 279 patients were identified with VAP based on one or more of the three criteria. Only 25.4% met NHSN criteria, while 88.2% met NTDB and 76.3% met Cx criteria. Only 20.1% patients met all three criteria, while 48.4% met two of three criteria and 30.8% met a single criterion. When NHSN criteria are compared to our “gold standard” Cx criteria, the NHSN shows a high specificity of 92.5% and low sensitivity of 28.2%. The PPV is 84.5% but the NPV is low at 47.1%. Kappa statistic of 0.18 shows poor agreement between the two criteria. The NTDB compared to the Cx criteria has a low specificity at 57.8% and high sensitivity of 86.4%. The PPV and NPV are both 74% and the kappa statistic of .47 shows fair agreement between the criteria. Conclusions: Within our institution, there is a wide range of VAP rates reported depending on which set of criteria are used. VAP is likely underreported to the NHSN as only about 1 in 4 of all VAPs identified met NHSN criteria. Alternatively, VAP may be over-reported to TQIP, as 88.2% met our NTDB criteria. Agreement of the NHSN with our institutional criteria was poor, but agreement between NTDB and our institutional criteria was much higher. Despite a NTDB definition of pneumonia, the lack of defined criteria for VAP may result in variable reporting to TQIP. Further work should be done to establish a consensus VAP definition.